Re: [Qemu-devel] [PATCH 3/4] ppc: booke206: use MAV=2.0 TSIZE definition, fix 4G pages

[Andreas Färber]

Am 07.05.2012 18:28, schrieb Alexander Graf:
> Hi Fabien,
> 
> Could you please elaborate a bit on the case that broke for you with these? 
> The patches shouldn't change any guest facing behavior :o.
> 
> 
> Alex
> 
> On 07.05.2012, at 17:47, Fabien Chouteau wrote:
> 
>> Hi Scott,
>>
>> I'm a little bit late, but this patch is not compatible with e500.
>>
>> In fact all the modification breaks e500v2 MMU support. What kind PPC
>> core are you working on?
>>
>> Regards,
>>
>> On 07/08/2011 01:44 AM, Scott Wood wrote:
>>> This definition is backward compatible with MAV=1.0 as long as
>>> the guest does not set reserved bits in MAS1/MAS4.
>>>
>>> Also, fix the shift in booke206_tlb_to_page_size -- it's the base
>>> that should be able to hold a 4G page size, not the shift count.
>>>
>>> Signed-off-by: Scott Wood <address@hidden>
>>> ---
>>> Unchanged in patchset v2
>>>
>>> hw/ppce500_mpc8544ds.c |    2 +-
>>> target-ppc/cpu.h       |    4 ++--
>>> target-ppc/helper.c    |    5 +++--
>>> 3 files changed, 6 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/hw/ppce500_mpc8544ds.c b/hw/ppce500_mpc8544ds.c
>>> index 3626e26..1aed612 100644
>>> --- a/hw/ppce500_mpc8544ds.c
>>> +++ b/hw/ppce500_mpc8544ds.c
>>> @@ -187,7 +187,7 @@ out:
>>> /* Create -kernel TLB entries for BookE, linearly spanning 256MB.  */
>>> static inline target_phys_addr_t booke206_page_size_to_tlb(uint64_t size)
>>> {
>>> -    return (ffs(size >> 10) - 1) >> 1;
>>> +    return ffs(size >> 10) - 1;
>>> }
>>>
>>> static void mmubooke_create_initial_mapping(CPUState *env,
>>> diff --git a/target-ppc/cpu.h b/target-ppc/cpu.h
>>> index f8bf2b1..9cf8327 100644
>>> --- a/target-ppc/cpu.h
>>> +++ b/target-ppc/cpu.h
>>> @@ -654,8 +654,8 @@ enum {
>>> #define MAS0_ATSEL_TLB     0
>>> #define MAS0_ATSEL_LRAT    MAS0_ATSEL
>>>
>>> -#define MAS1_TSIZE_SHIFT   8
>>> -#define MAS1_TSIZE_MASK    (0xf << MAS1_TSIZE_SHIFT)
>>> +#define MAS1_TSIZE_SHIFT   7
>>> +#define MAS1_TSIZE_MASK    (0x1f << MAS1_TSIZE_SHIFT)
>>>
>>> #define MAS1_TS_SHIFT      12
>>> #define MAS1_TS            (1 << MAS1_TS_SHIFT)
>>> diff --git a/target-ppc/helper.c b/target-ppc/helper.c
>>> index 176128a..892c6e3 100644
>>> --- a/target-ppc/helper.c
>>> +++ b/target-ppc/helper.c
>>> @@ -1293,7 +1293,7 @@ target_phys_addr_t booke206_tlb_to_page_size(CPUState 
>>> *env, ppcmas_tlb_t *tlb)
>>> {
>>>     uint32_t tlbncfg;
>>>     int tlbn = booke206_tlbm_to_tlbn(env, tlb);
>>> -    target_phys_addr_t tlbm_size;
>>> +    int tlbm_size;
>>>
>>>     tlbncfg = env->spr[SPR_BOOKE_TLB0CFG + tlbn];
>>>
>>> @@ -1301,9 +1301,10 @@ target_phys_addr_t 
>>> booke206_tlb_to_page_size(CPUState *env, ppcmas_tlb_t *tlb)
>>>         tlbm_size = (tlb->mas1 & MAS1_TSIZE_MASK) >> MAS1_TSIZE_SHIFT;
>>>     } else {
>>>         tlbm_size = (tlbncfg & TLBnCFG_MINSIZE) >> TLBnCFG_MINSIZE_SHIFT;
>>> +        tlbm_size <<= 1;
>>>     }
>>>
>>> -    return (1 << (tlbm_size << 1)) << 10;
>>> +    return 1024ULL << tlbm_size;

Here the page size changes, doesn't it? The << 1 shift is only happening
in the else branch whereas it was always done before.

Andreas

>>> }
>>>
>>> /* TLB check function for MAS based SoftTLBs */

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg